Learn R Programming

blin (version 0.0.1)

Bipartite Longitudinal Influence Network (BLIN) Estimation

Description

Estimate influence networks from longitudinal bipartite relational data, where the longitudinal relations are continuous. The outputs are estimates of weighted influence networks among each actor type in the data set. The generative model is the Bipartite Longitudinal Influence Network (BLIN) model, a linear autoregressive model for these type of data. The supporting paper is ``Inferring Influence Networks from Longitudinal Bipartite Relational Data'', which is in preparation by the same authors. The model may be estimated using maximum likelihood methods and Bayesian methods. For more detail on methods, see Marrs et. al. .

Copy Link

Version

Install

install.packages('blin')

Monthly Downloads

10

Version

0.0.1

License

MIT + file LICENSE

Maintainer

Frank W. Marrs

Last Published

September 21st, 2018

Functions in blin (0.0.1)

Generate data from the continuous BLIN model

model.matrix.blin

model.matrix S3 generic for class blin

vcov S3 generic for class blin

Online forum dataset

fit_MLE_array_additive

Fit full BLIN model D is matrix of lagged Y (covariates) Y is oservations X is covariates for multiple time periods Returns A, B, beta, betaOLS, Yhat, 2LL, 2LLinit

update_MLE_asymmetric

D is matrix of lagged Y (covariates) Y is observations

Fit reduced rank BLIN model D is matrix of lagged Y (covariates) Y is oservations X is covariates for multiple time periods Returns A, B, beta, betaOLS, Yhat, 2LL, 2LLinit

Plot S3 generic for class blin

Compute generalized inverse of X^T X, for X

Lag Y array to get autoregressive array D

Matricize an array

Print S3 generic for class blin

print.summary.blin

Print S3 generic for class summary.blin

Summary S3 generic for class blin

update_MLE_additive

Update for block coordinate descent of full BLIN model D is matrix of lagged Y (covariates) Y is observations

Check inputs for reduced rank fit of BLIN model

Check whether there is a unique solution to the BLIN model

Multiply an array by a matrix along a given mode, by

additive_regression

Fit sparse BLIN model D is matrix of lagged Y (covariates) Y is oservations X is covariates for multiple time periods Returns A, B, beta, Yhat

Find diagonal indices of k-mode array, where "diagonal" is i=j for first two indices of input array Y

Coef S3 generic for class blin

Build the BLIN design matrix

Check inputs to blin_fit()

Estimate the BLIN model using maximum likelihood estimator